AI Agent Operational Lift for Uc San Diego Library in La Jolla, California

Why AI matters at this scale

UC San Diego Library is a mid-sized academic library serving a top-tier research university with over 40,000 students and faculty. It manages millions of volumes, extensive digital collections, special archives, and data services. With 201–500 employees, the library operates at a scale where manual processes strain resources, yet it lacks the vast budgets of mega-libraries. AI offers a force multiplier—automating repetitive tasks, enhancing user experience, and unlocking hidden value in collections.

What UC San Diego Library does

The library provides access to scholarly resources, research support, instruction, and preservation of cultural heritage. Its digital initiatives include institutional repositories, data curation, and digitized special collections. The library’s mission is to advance knowledge creation and student success, making it a critical hub on campus. However, like many academic libraries, it faces rising expectations for instant, personalized services while managing flat or shrinking budgets.

Why AI is a strategic fit

At this size, the library has enough data and infrastructure to benefit from AI without the complexity of a massive enterprise. It already uses systems like Ex Libris Alma and OCLC, which can integrate AI plugins. AI can address three core pain points: (1) metadata backlogs that delay discoverability, (2) growing demand for research assistance that outpaces staff capacity, and (3) underutilized digital archives that lack full-text searchability. By adopting AI, the library can reallocate staff to higher-value work like instruction and outreach.

Three concrete AI opportunities with ROI

Automated metadata generation – Using NLP to create MARC records, summaries, and keywords can cut cataloging time by 50–70%. For a library adding tens of thousands of items yearly, this saves thousands of staff hours, translating to over $200,000 in annual labor cost avoidance. Faster cataloging also means resources are available to users sooner, boosting usage.

AI-powered search and discovery – Replacing keyword-based search with semantic vector search improves relevance and serendipity. Early adopters report a 20–30% increase in full-text downloads and user satisfaction. For UC San Diego, this could mean higher engagement with licensed databases and special collections, justifying collection investments.

Research assistant chatbot – A GPT-based chatbot handling 40% of routine reference questions (e.g., “How do I access this journal?”) frees librarians for complex consultations. With 50,000+ annual reference interactions, even a 30% deflection rate saves 3,000+ staff hours, valued at $150,000+. It also provides 24/7 support, meeting student expectations.

Deployment risks specific to this size band

Mid-sized libraries face unique challenges: limited in-house AI expertise, tight budgets for experimentation, and the need to integrate with legacy library systems. Data privacy is paramount—user borrowing and search data must be anonymized to comply with FERPA and library ethics. Staff resistance is real; change management and upskilling are essential. Starting with low-risk pilots, leveraging vendor partnerships, and measuring clear outcomes can mitigate these risks. The library’s existing culture of innovation and campus IT support provides a strong foundation for responsible AI adoption.